Impacts of Group Prenatal Care on Birth Outcomes: Evidence from South Carolina's CenteringPregnancy Expansion

Wednesday, June 13, 2018: 10:20 AM
1034 - First Floor (Rollins School of Public Health)

Presenter: Jessica Smith

Co-Authors: Emily Heberlein; Carla Willis; Amy Crockett; Sarah Covington-Kolb

Discussant: Nicardo Mcinnis

CenteringPregnancy is a model of group prenatal care that has been associated with improved newborn and maternal health outcomes, including potential reductions in racial disparities. Following the recommendations of the South Carolina Birth Outcomes Initiative (SC BOI) Health Disparities Workgroup, the South Carolina Department of Health and Human Services (SC DHHS) began investing in the expansion of CenteringPregnancy in health care practices across the state in January 2013.

South Carolina has historically high rates of poor birth outcomes, including preterm birth and low birthweight, and significant disparities by race. In order to understand the impact of this innovative state-wide expansion, our study compares outcomes for women who participated in CenteringPregnancy to women who received traditional, individual prenatal care using birth certificate data. This study includes outcomes for babies born between January 2013 and May 2017. While expansion practices offered CenteringPregnancy to women with all types of insurance coverage (including self-pay/uninsured), our work focuses on women enrolled in Medicaid. This retrospective cohort analysis uses birth certificate data from the South Carolina Revenue and Fiscal Affairs Office (RFA) to evaluate the impact of CenteringPregnancy on six birth outcomes (preterm birth, low birthweight, neonatal intensive care unit (NICU) admissions, cesarean delivery, breastfeeding, and gestational diabetes) across 13 sites.

CenteringPregnancy practices varied in total size, volume of groups, and the proportion of their patients enrolled in Medicaid. Because randomizing patients to CenteringPregnancy or individual care was not possible in this scale-up intervention, we used available characteristics of women to develop propensity scores to account for differences between the groups who chose CenteringPregnancy or individual care. Our analysis employs a preferential-within matching technique, matching CenteringPregnancy patients to other patients with similar propensity scores within each practice site. If a similar individual care patient was not found within the same site, a similar individual care patient was matched from another site. This technique accounts for the nested nature of the data (where women received prenatal care at 13 sites). The impact of CenteringPregnancy on each outcome was evaluated using logistic regression models with the matched CenteringPregnancy and individual care cohorts. Each analysis was repeated for three groups, reflecting a progression in comparisons of prenatal care and CenteringPregnancy “dose” or “participation”.

Participation in CenteringPregnancy positively impacted birth outcomes for women enrolled in Medicaid. Across prenatal care participation levels, CenteringPregnancy patients consistently had better outcomes. Centering participation reduced the prevalence of births resulting in NICU admissions by 2.8 percentage points (p<0.001), decreased low birthweight deliveries by 2.2 percentage points (p<0.001), and increased breastfeeding rates by 9 percentage points (p<0.001). Most differences in birth outcomes were statistically significant, with the exception of preterm birth (when comparing any Centering participation to all other women), cesarean delivery and gestational diabetes outcomes. The impact of CenteringPregnancy participation was even greater among Black women and women attending five or more Centering visits.